Add python_abi output and refactor #1
Conversation
isuruf
commented
Jan 8, 2026
isuruf
commented
- Remove clang-19 since we are not building the experimental jit
- Remove ld_impl because we are not hacking the makefiles
- Add site_packages_path
- Make recipe.yaml identical with just the variant name difference
- Add a generate.sh to generate recipe.yaml from default variant
isuruf
force-pushed
the
tsan
branch
11 times, most recently
from
9a4b234 to
d810a1e
Compare
- Remove clang-19 since we are not building the experimental jit - Remove ld_impl because we are not hacking the makefiles - Add site_packages_path - Make recipe.yaml identical with just the variant name difference - Add a generate.sh to generate recipe.yaml from default variant - Add a runtime requirement of libsanitizer for gcc
|
Ping @lucascolley in case you're curious how Isuru sorted out the |
Tools/pixi-packages/generate.sh
Outdated
| if [[ "${name}" == *free-threading* ]]; then | ||
| tag="cp35t" | ||
| else | ||
| tag="cp35" |
There was a problem hiding this comment.
shouldn't this be cp315? Or was this choice on purpose?
|
nice :) and on the Pixi side we have progress towards being able to use a single recipe with |
|
Can we have one recipe and pass a variant using |
| name: python_abi | ||
| version: ${{ version }} | ||
| build: | ||
| string: "0_cp315" |
There was a problem hiding this comment.
@lucascolley I had to change this from 0_${{ abi_tag }} because pixi cannot handle this correctly. Pixi would expect a build string of 0_ and complaint that it found 0_cp315. Probably an issue with conda/outputs thing in pixi build.
There was a problem hiding this comment.
sounds right, a bug report would be perfect
That would indeed be more simple once the variant-sensitivity is added for Pixi workspace dependencies |
|
The readme probably needs an update too: https://github.com/python/cpython/tree/main/Tools/pixi-packages#maintenance |
|
@crusaderky when you’re able to come back to this, let us know if this looks good. |
3 participants
